entrypoint.sh

#!/bin/bash

# 1. 后台启动 Ollama 服务
echo "Starting Ollama server..."
ollama serve &

# 2. 等待服务完全启动 (避免立即拉取模型导致失败)
until curl -s localhost:7860 > /dev/null; do
  echo "Waiting for Ollama to start..."
  sleep 2
done

# 3. 预下载模型 (如果不下载，第一次 API 调用会非常慢)
echo "Downloading model: hf.co/apothic/bonsai-8B-1bit-turboquant..."
ollama run hf.co/apothic/bonsai-8B-1bit-turboquant "hello"

# 4. 保持进程运行，防止 Docker 容器退出
echo "Model ready! Space is running on port 7860."
wait